Algorithm Research On State Parameter Estimation For Hub Motor Drive Vehicles

The wheel torque of hub motor drive vehicle is independently controllable and responsive,which is conducive to the implementation of vehicle dynamics control.However,the critical parameters such as longitudinal speed and road surface adhesion coefficient required for dynamics control are difficult to rely on for basic vehicles sensor acquisition.For this reason,the paper developed a real-time estimation algorithm for the state parameters of hub motor drive vehicles.The main research work is as follows:1)Establishment of simulation model of hub motor driven electric vehicle.Relying on the real vehicle platform of the research group,the 7 degree of freedom vehicle simulation model based on the Dugoff tire model is established by Matlab / Simulink and Carsim software platform,and the model is verified by preliminary simulation.2)Estimation algorithm design of road adhesion coefficient.An unscented Kalman filter estimation algorithm based on fading memory is proposed.By fading memory filtering,the observed data can be effectively using,and the system state variables are weighted in real time to improve the accuracy of parameter estimation.On the other hand,combined with the advantage of in-wheel motor driven vehicle driving torque measurable,it is proposed that the longitudinal force obtained by driving wheel model is used to corrected the longitudinal force of the Dugoff tire model by feedback,so as to improve the accuracy of tire force input required by the estimation algorithm,and indirectly improve the accuracy of road adhesion coefficient estimation.3)Estimation algorithm design of vehicle driving state parameters.Considering the model error and the strong coupling between the observed parameters that includes longitudinal and lateral speed,yaw rate and side slip angle,an adaptive unscented Kalman filter estimation algorithm based on fuzzy logic correction is proposed,the fuzzy logic system can correct the system process noise and measurement noise in real-time,thereby improving the accuracy and condition adaptability of parameter estimation.In addition,the road slope parameters are integrated into the system state variables to improve the algorithm utilization.4)Integrated design of observation algorithm and real vehicle test.Through the integrated design of the above two different estimation algorithms,the mutual interaction and the mutual correction between the two estimated parameters are achieved to achieve the integrity of the system.Then,the simulation verification of integrated algorithm through simulation platform,and combined with the real vehicle platform of the research group,and the real vehicle test of the integrated algorithm under different working conditions was completed.Simulation and real vehicle verification test results show that the unscented Kalman filter algorithm based on fading memory and the adaptive unscented Kalman filter algorithm based on fuzzy logic correction developed by the thesis improve the estimation of the road surface adhesion coefficient and vehicle driving state parameters.Accuracy and stability can guarantee the performance requirements of the required parameters of the vehicle's dynamics control.It has important theoretical guidance and practical application value.